@vanderwal a few months ago i ran a WNYC audio segment through whisper (maybe in a python REPL) and got a good transcript (maybe i can find those notes?)